Most B-cell lymphomas arise from germinal center (GC) B-cells, which form transiently after T-cell dependent antigen stimulation. GC B-cells undergo massive proliferation and genomic instability occurring as a byproduct of immunoglobulin somatic hypermutation, which puts them in danger of malignant transformation. The phenotypic shift from quiescent nave B-cells to proliferative and unstable GC B-cells is massive, rapid, and involves differential expression of thousands of genes. GC B-cells are evanescent, and quickly undergo terminal differentiation to plasma cells (or undergo apoptosis) after antigen presentation. The most common B-cell lymphomas (DLBCL and FL) in essence are GC B-cells that have continued to aberrantly persist and fail to undergo terminal differentiation. We are interested in how these dramatic changes in phenotypes occur, and how this process can be corrupted to cause lymphoma. To understand the mechanistic basis of the GC B-cell phenotype we performed genome-wide chromosomal conformation capture (Hi-C, 4C) along with ChIP-seq for histone marks, cohesin and TFs at different timepoints during B-cell development. We observed truly massive shifts in chromosomal architecture in GC B-cells including but not limited to i) increased promoter connectivity, ii) formation of novel enhancer loops, iii) 5? to 3? gene looping, iv) merging of discrete boundary delimited gene neighborhoods to form larger gene ?cities resulting in de novo epigenetic coordination between genes formally isolated from one another, and v) establishment of GC B-cell specific locus control regions (LCRs) that control hundreds of GC B-cell gene enhancers (Bunting et. al. Immunity 2016). Strikingly, all of these architectural changes were tightly associated with cohesin complex redistribution and notably, we observed recurrent somatic mutation or deletion of the cohesin unloading protein PDS5B in public lymphoma genomic profiling datasets. Our pilot studies suggest that PDS5B regulates genes involved in exiting the GC reaction and terminal differentiation. Preliminary experiments in PDS5b knockout or point mutant mice, point to disruption of GC dynamics and blockade of GC exit. Based on these considerations we hypothesize that PDS5B is required to unload the GC specific cohesin distribution patter state so that the GC B-cell transcriptional program can be extinguished and allow for a different configuration that favors plasma cell differentiation. We predict that specific signals received from GC T-cells in the GC light zone directly induce PDS5B-dependent cohesin redistribution. We propose that genetic lesions of PDS5B cause the genome to become architecturally stuck in the GC configuration thus blocking epigenetic reprogramming required for terminal differentiation and leading to malignant transformation. We hypothesize that cohesin blockade may be nonetheless reversible and targetable by drugs that can erase GC/lymphoma epigenetic programming. This proposal will thus define the role and mechanism of action of dynamic cohesin complex remodeling in the humoral immune response and lymphomagenesis, and develop novel cohesin therapy approaches.